Overlay (Tracy, Leon) GlastRelease v15r50p6, made last week, has first version of overlay code, but old (original) triggerAlg code. This means overlay data is being considered by the trigger, not what is desired. New plan is to not overlay TKR/ACD/CAL digis until after TriggerAlg runs. Lots of gotchas in implementation: for example, In periodic triggers CAL and ACD data are not zero-suppressed. This has especially weird consequences for ACD, such as ostensible hits with energy < 0. TBD: when to do the zero-suppression. Should it be done to overlay files before merging? Would make these files significantly smaller. (Leon) is investigating ACD bugginess w.r.t. display, which may shed some light more generally. (Richard) What about 4-range read-out? (Tracy) Yes, another issue to investigate. (Philippe) So is it the case that we don't know how to reconstruct these ACD event (with no suppression) in any context?
Interacting alphas (Richard) Lucca and Francesco are ready for a tag including the interacting alpha code. The tag should be on the same branch as LPM, which is in tag v15r51p1.
FSSC Report: (Eric W., then John) Things are pretty quiet: just testing and working through bugs on various platforms, such as with celestialSources and some other packages on Linux RHEL4 64-bit. RM has the same problem on this platform.
ScienceTools: See the report.
Documentation: (Chuck) sends this summary:
Short report, but a lot happening. Last newsletter of the year is on the streets, and the restructuring of the workbook's top-level navbars is going well. I intend to post the changes to the web later today, or early tomorrow. That will enable me to make a new tag, so Berrie can update the mirror site before year's end. Jim Chiang also updated the Science Tools page: 'Sources Available to gtobssim and Gleam'. And Masa's still working hard on updating the pulsar tutorials but, in fact, has not given the go ahead yet to post them to the web.And..., that's the latest from WB World for this morning.
(Richard) The restructuring is to address the shift in interest (for the majority of customers) from GlastRelease to ScienceTools.
SCons RM (Navid) The SCons RM is now making release builds for Science Tools as well as HEAD and LATEST builds. RHEL4 32-bit looks good; 64-bit has problems (as noted above in FSSC report). The RHEL5 MySQL issue has been resolved (proper client library is installed) so that RM can run on the these machines. Macs do not yet have lsf.
GR and Externals (Emmanuel) GR v15r4p9 is ready for SCons: all SConscript files have been updated and other, non-SCons-specific files needing patches have been modified, committed and tagged according to a suitable convention. (Richard) How about getting these changes integrated into HEAD? (Heather) would like to get package owners involved.
(Emmanuel) The cfitsio patch fixing problems seen with large files is available for RHEL3, RHEL4, Mac and Windows.
GoGui et. al. (Joanne) No real news on GoGui proper since last week, when supersede (aka override) support made its appearance in version 0.9.1. [One bit of GoGui news since the meeting: Emmanuel's problem has been diagnosed. GoGui crashes when SCons is invoked if the SCons command path hasn't been established properly from the Utilities item in the Options menu. Ideally GoGui should catch the error and put up a message rather than crashing, but so far I haven't found a way. The workaround is clear, however: make sure your definition for SCons really does point to the a working instance of SCons. ]
The command-line SCons-aware tagger discussed last week, going by the name of stag, is written, documented, and available for download via ftp
Conversion of remaining packages in CHS to SCons made significant progress — modified code in the MootBuild class of package MOOT using the new interface compiles — but as of the end of last week was on hold until Owen finished making changes to the new flight software packages RIM and QCFG which will replace the old LATC package. [Owen released new versions of the flight packages late Tuesday.]
Concerning better SCons Windows support: the work done to date is for VS 2005 rather than 2003. There are significant differences in format of solution and project files between these two versions. Sentiment among Windows users is that we should push ahead to VS 2005 or even VS 2008, even though they are not yet formally supported: we'll need builds of external libraries (VS 2003-built libraries cannot be used with the newer platforms). There is no significant difference in project/solution file format between 2005 and 2008.
previous | minutes index | next |